Cross Domain Few-Shot Learning via Meta Adversarial Training

TL;DR

This paper introduces a novel meta adversarial training framework for cross-domain few-shot relation classification, enabling models trained on source domains to adapt effectively to target domains without target data during training.

Contribution

It proposes a meta-based adversarial training framework that allows domain adaptation in few-shot learning using only source domain data for training.

Findings

The model effectively adapts to target domain data.

Empirical results confirm improved cross-domain classification performance.

The approach outperforms baseline methods in experiments.

Abstract

Few-shot relation classification (RC) is one of the critical problems in machine learning. Current research merely focuses on the set-ups that both training and testing are from the same domain. However, in practice, this assumption is not always guaranteed. In this study, we present a novel model that takes into consideration the afore-mentioned cross-domain situation. Not like previous models, we only use the source domain data to train the prototypical networks and test the model on target domain data. A meta-based adversarial training framework (MBATF) is proposed to fine-tune the trained networks for adapting to data from the target domain. Empirical studies confirm the effectiveness of the proposed model.